Abstract: In the complex tunnel environments of coal mines, traditional centralized vehicle dispatch systems suffer from inherent problems of high deployment costs and significant response delays due to reliance on standalone controllers; concurrently, existing positioning technologies fail to achieve continuous cross-base-station vehicle tracking while lacking real-time monitoring capabilities for violations such as speeding and red-light running. To address these challenges, a distributed intelligent traffic control system based on joint scheduling units was constructed, the study designs a master base station (integrated with intelligent scheduling algorithms) and multi-slave base station collaborative architecture; the master station actively polls ultra-wideband (UWB) vehicle positioning data from slave stations via TCP/IP protocols, dynamically generating traffic light control commands based on configurable locking and unlocking distance ranges to achieve centralized-controller-free distributed decision-making; supports modular deployment configurations including “master-slave pairs” for straight sections, “master-dual-slave” for T-junctions, and “master-triple-slave” for crossroads, while significantly reducing hardware construction costs through base station reuse technology enabling single stations to serve dual joint scheduling units; the study develops a seamless roaming positioning mechanism utilizing the maximum ranging distance cross-coverage, allowing vehicle cards to synchronously scan neighboring base stations and triggering positioning base station handover when consecutively entering a new station ranging zone three times, thereby eliminating trajectory voids exceeding 10 seconds in traditional solutions; the study establishes a violation recognition model that calculates real-time vehicle speed via UWB ranging displacement differentiation combined with traffic light state machines for dynamic red-light violation detection. Field validation at Shanxi Datong Coal Mine demonstrates that: traffic light control accuracy exceeding 99.3%, speeding detection rate no less than 98.2%, red-light violation recognition reaching 99.1%, and positioning handover time stabilized within 2 seconds in multi-base-station scenarios, fully verifying the excellent performance in dispatch control accuracy, safety supervision comprehensiveness, and positioning continuity of the system.